00:14 <davidc__> ysionneau: the design files they published were for a two layer board

00:15 <davidc__> the ground plane was on the bottom, and very sliced up by traces

00:15 <davidc__> so there were some really suboptimal return paths for current.

00:16 <davidc__> ysionneau: I fixed the issue by scraping off the soldermask over the ground plane on both sides of the cut in the plane

00:16 <davidc__> and then soldering a wire to bridge the cut.

07:14 <rjo_> davidc__: nice. thanks for that bit of info. ist the break on the fron or on the back?

07:14 <rjo_> s/ist/is/ and s/fron/front/

07:31 <davidc__> rjo_: backside, all the traces interrupting the ground plane

07:31 <davidc__> (the design files we had were 2 layer only)

07:31 <davidc__> rjo_: I just found every place the plane was split with a trace, and jumped it

07:35 <rjo_> davidc__: looks like the cuts are mostly horizontal.

07:38 <davidc__> rjo_: yeah. but theres nothing on the top either :)

07:38 <rjo_> davidc__: ok. but you did not bridge every single trace on the back. a couple judiciously placed vertical bridges should do it, right?

07:39 <davidc__> rjo_: yeah, sorry - I meant "every place where the ground plane was split creating a large-loop-area" or something along those lines

07:40 <rjo_> davidc__: ack. thanks for pointing this out! ist jack aware of this?

07:40 <davidc__> rjo_: dunno, no idea

07:41 <davidc__> rjo_: I assumed it had been fixed for the production files and that the open-source design files were just out of date

07:41 <davidc__> and didn't think much more of it

07:41 <davidc__> rjo_: I don't actually have one of those boards, just helped the guy doing the build at the hackspace figure out why the boards had flashing issues

07:43 <davidc__> rjo_: anyhow, let me know if that fixes it. I'm curious to know if its in the production units as well

08:08 <rjo_> davidc__: the current files are the same as what is being currently sold.

08:10 <rjo_> davidc__: it is marginally working/non-working depending on the (quantum) phase of the moon.

08:13 <davidc__> fwiw, the exact symptoms on my test board were random verification failures during programming,

08:14 <davidc__> and fixing the ground plane made programming reliable (couldnt make it fail after that)

08:32 <rjo_> davidc__: was that ground bounce on erase or write or read?

16:57 <davidc__> rjo: every sck edge